Median Rate Ratio Research Articles

Abstract 335: Cardiac Rehabilitation Referral After Acute Myocardial Infarction: Variation in Referral Rates and Association with Core Components of Risk Factor Management

Background: Although currently an endorsed performance measure of quality, cardiac rehabilitation (CR) referral rates remain quite low. Established guidelines describe recommended core components of CR programs for risk factor management: nutritional counseling; physical activity counseling; exercise training; smoking cessation; and weight, hypertension, lipid, diabetes, and psychosocial management. We sought to determine the variability in hospital referral rates for CR and assess whether patients are more likely to be referred who have risk factors that are intervention targets of the core components of CR. Methods: We assessed referral to outpatient CR among 3632 MI patients from 22 US hospitals in the TRIUMPH study who were discharged home with risk factor data available. Variation in CR referral rates across hospitals was quantified by the median rate ratio (MRR). The prediction of CR referral by 8 individual risk factors that are intervention targets of CR was assessed using hierarchical modified Poisson regression, adjusting for study site and 16 demographic and clinical patient characteristics. Results: Upon hospital discharge, only 39.1% (n=1420) of patients were referred to CR. Referral rates varied greatly across sites, ranging from 2.6% to 88.2%. After adjusting for potentially confounding patient factors, the MRR was 2.47 (95% CI 1.67, 3.23), indicating an average 2.5 times greater likelihood of 2 hypothetically identical patients being referred to CR upon discharge from 1 random hospital as compared with another. Only 1 of 8 risk factors modifiable by CR core components was associated with increased referral (see figure), with overweight patients (BMI > 25 or waist circumference > 40 in. for men, > 35 in. for women) being more likely to be referred than non-overweight patients (RR 1.09; 95% CI 1.03, 1.15; p<0.01). Conclusion: Among patients discharged home after MI hospitalization, there are marked differences in outpatient CR referral rates by study site, but relatively little association with patient risk factors that the core components of CR programs are specifically designed to modify. This suggests that site-specific factors strongly influence CR referral and site-level interventions to improve referral are likely to most effectively improve compliance with this performance measure.

PCN16 - Hospital Center Effect For Laparoscopic Colectomy Among Elderly Stage I-Iii Colon Cancer Patients

OBJECTIVE: To investigate hospital-level variation in short-term laparoscopic colectomy outcomes among stage I-III elderly colon cancer patients. BACKGROUND: Surgical outcomes are associated with patient and surgeon characteristics. If outcomes are also impacted by the hospital where the surgery occurs, there is a hospital center effect (HCE). METHODS: Surveillance, Epidemiology, and End Results (SEER)-Medicare data was used to identify stage I-III colon cancer patients treated with laparoscopic colectomies. Multilevel regressions were utilized to study potential HCE for length of stay (LOS), 30-day rehospitalization, and in-hospital mortality, adjusting for patient, surgeon, and hospital-level characteristics. To quantify HCE, we calculated the median instantaneous rate ratio (MIRR) for LOS and median odds ratio (MOR) for in-hospital mortality and 30-day rehospitalization. Sensitivity analyses were conducted for high volume/medical school affiliated hospitals and colorectal surgeons. RESULTS: The multilevel analyses based on 4617 patients from 465 hospitals documented statistically significant HCEs for LOS (MIRR = 1.35; P < 0.001) and in-hospital mortality (MOR = 1.69; P = 0.032), but no HCE for 30-day rehospitalization. Sensitivity analyses confirmed our findings. HCE was significant for LOS in all sensitivity analyses and was significant for in-hospital mortality for high volume/medical school affiliated hospitals. CONCLUSIONS: HCE exists for LOS and in-hospital mortality of laparoscopic colectomy, which suggests that the choice of hospital affects outcomes independently of other confounding variables. Reducing the variation in outcomes associated with HCE may improve the quality of cancer care.

Abstract 75: Practice-level Variation in use of Recommended Medications among Outpatients with Heart Failure: Insights from the NCDR PINNACLE Program

Background: Beta-blockers (BB) and angiotensin converting enzyme inhibitors (ACE)/angiotensin receptor blockers (ARB) are guideline recommended therapies for ambulatory heart failure patients with reduced left ventricular ejection fraction (HFREF). Real-world use of these therapies is suboptimal. Whether variations in treatment are dominated by practice-level or patient-level factors is unknown. Methods: Practices participating in the NCDR PINNACLE program, a national office-based cardiac quality improvement registry, between July 2008 and December 2010 were evaluated. Rates of treatment with BB and ACE/ARB were evaluated in patients with HFREF (EF ≤40%) and no documented contraindication. Multivariable hierarchical relative risk regression models, including demographics, insurance status, comorbidities and a random effect for practice were used to determine 1) patient-level and 2) practice-level variation in treatment rates. To quantify practice-level variation, the median rate ratio (MRR) was calculated, which estimates the typical rate ratio between two randomly selected practices for patients with identical covariates. In general, MRRs ≥ 1.2 indicate significant variation by practice. The MRR is always &gt;1.0 but can be compared in magnitude to patient-level risks. Results: We studied 12384 patients in 45 practices. The mean practice rate for BB treatment was 87% (IQR 83%-95%; range: 43%-100%), and the mean practice rate of ACE/ARB treatment was 90% (IQR 75%-88%; range: 18%-100%). The MRR was 1.09 for BB and 1.16 for ACE/ARB therapy. For both BB and ACE/ARB, the adjusted MRR for site level variation was larger than the rate ratio for other patient factors. (Table) Conclusions: Although rates of BB and ACE/ARB treatment among outpatients with HFREF are high, clinically meaningful variation by practice is present and explains a larger amount of the observed variance than any patient characteristic. This suggests that addressing practice-level factors represents an important opportunity to improve the use of evidence-based HF therapy.

Abstract 257: Practice Variation is a Significant Contributor to Secondary Prevention Medication Use: Insights from the NCDR PINNACLE Program

Background: Secondary prevention medications for eligible CAD patients include antiplatelets, lipid lowering agents, beta-blockers (BB) and angiotensin converting enzyme inhibitors (ACEI) or angiotensin receptor blockers (ARB). However, use of these therapies is variable. The contribution of practice-level factors to variation in medication use is unknown. In addition, the contribution of practice-level factors relative to patient factors is also unknown. Methods: We evaluated CAD patients in cardiology practices participating in the NCDR PINNACLE program, a national office-based cardiac quality improvement registry, between July 2008 and December 2010. Patients eligible for treatment were grouped by Class I indications for therapy: antiplatelets and lipid lowering agents in all, BBs in post-MI patients, and ACEIs/ARBs in those with diabetes or with reduced left ventricular ejection fraction (≤40%). Mean practice rates of therapy use were calculated. Next, hierarchical regression models assessed the effect of practice on therapy use, adjusted for patient factors, using the median rate ratio (MRR). The MRR compares the likelihood of treatment for identical patients from two randomly selected practices. In general, MRRs ≥ 1.2 indicate significant variation by practice. Finally, the magnitude of contribution of patient factors, expressed in ORs, to variation in therapy use was assessed. Results: Our cohort consisted of 277,526 patients in 62 practices. The mean practice rate was 80.8% for antiplatelets (range 41.2–100%), 79.5% for lipid lowering agents (range 33.9–100%), 77.6% for BBs (range 35.2–100%), and 73.8% for ACEI/ARBs (range 39.1–100%). Adjusted MRR between practices was 1.19 for antiplatelets, 1.19 for lipid lowering agents, 1.20 for BBs, and 1.19 for ACEIs/ARBs. Almost all patient-level factors were smaller in magnitude than the MRR for practice-level variation (see Table). Conclusions: Among practices participating in the PINNACLE program, significant variation in guideline recommended treatments among CAD outpatients exists. Practice setting, independent of patient factors, was associated with variation in treatment and was similar in magnitude to most patient factors. Our findings suggest that practice setting is an important contributor to variability in secondary prevention medication use. Accordingly, to improve the use of evidence-based CAD therapies, quality improvement efforts targeting practice-level factors should be evaluated.

Abstract 274: Patterns Of Statin Initiation, Intensification And Maximization Among Patients Hospitalized With An Acute Myocardial Infarction. Insights From The TRIUMPH Prospective Cohort Study.

Background: Intensive statin therapy after an acute myocardial infarction (AMI) has been shown, in multiple trials, to be superior to less intensive treatment. Several studies have documented rates of statin prescription at discharge, in any dose, as a quality performance measure. However, variations in hospitals’ rates of initiating, intensifying and maximizing statin therapy after AMI are unknown. Methods: We assessed statin therapy at admission and discharge among 4340 AMI patients from 24 US hospitals (4/05-12/08). Initiation was defined as statin prescription at discharge in patients not admitted on them. Intensification reflected an increase in dose or change to a higher potency statin at discharge. The use of atorvastatin 80 mg or rosuvastatin 40 mg at discharge were considered maximal statin therapy. Hierarchical modified Poisson models estimated the association of site with initiation in naïve patients, intensification in those who arrived on submaximal therapy, and discharge on maximal therapy among all patients, after adjustment for patient factors including LDL level. Site variation was explored with a median rate ratio (MRR), which estimates the average relative difference in risk ratios of 2 hypothetically identical patients treated at 2 different sites. Results: On admission, 34.8% (1510/4340) of patients were taking a statin and 2.6% (112/4340) were on maximal therapy. 86.9% (2411/2776) of statin naïve patients without a contraindication were prescribed a statin at discharge, with no variability across sites after adjustment for patient factors (MRR 1.02). Among patients who arrived on submaximal statin therapy, 26.3% (323/1230) had their statin therapy intensified at discharge with modest site variability (MRR 1.47). Among all patients without a contraindication, 21.0% (899/4271) were discharged on maximal therapy with substantial site variability (MRR 2.79). Conclusions: In a large, multicenter AMI cohort, nearly all patients were started on statins during hospitalization, with no variability across sites. However, rates of statin intensification and maximization were lower and varied substantially across hospitals. Given that more intense statin therapy is important in optimizing outcomes, evolving the existing performance measures to include the intensity of statin therapy may be beneficial in improving care.

Abstract P106: Variation in the Incidence of Hospital-Acquired Anemia During Admission with Acute Myocardial Infarction Across 57 US Hospitals

Background: New onset, hospital-acquired anemia (HAA) during acute myocardial infarction (AMI) may be related to hospital-based processes of care. HAA is associated with poor outcomes, but little is known about the extent of hospital variation in the incidence of HAA or the hospital characteristics associated with HAA. Methods: We studied 17,676 AMI patients not anemic at admission, defining moderate-severe HAA as a hemoglobin decline to &lt; 11 g/dl. Shrinkage estimates of moderate-severe HAA incidence were generated to account for low volume sites. Multivariable models were used to identify adjusted variation in moderate-severe HAA across hospitals, using median rate ratios (MRR - median value of the relative risk of moderate-severe HAA for two identical patients presenting to two randomly selected hospitals) and fit a separate model to test the association between hospital characteristics (# beds, region, urban/rural, teaching status) and moderate-severe HAA adjusting for patient factors. Results: Overall, 3,551 (20%) developed moderate-severe HAA. The incidence of moderate-severe HAA varied substantially across sites (Figure) and was significant after multivariable adjustment (MRR 1.3 [1.2-1.4]). The only site factors independently associated with moderate-severe HAA were teaching status (RR 0.7 [0.6-0.9] vs. non-teaching) and region (South vs. Midwest: RR 1.3 [1.0-1.5]). Conclusions: We found significant variation in the incidence of moderate-severe HAA and a lower risk of HAA in teaching hospitals. Further study of the relationship between HAA and specific processes of care is needed to identify actionable targets for quality improvement.

Abstract P340: Prevalence and Correlates of Smoking Cessation Pharmacotherapy in Hospitalized Smokers With Acute Myocardial Infarction

Background ACC/AHA guidelines recommend pharmacotherapy for smoking cessation in smokers with AMI. The aim of this study was to determine the prevalence of smoking cessation pharmacotherapy in hospitalized patients with AMI and to identify patient factors associated with prescription of these medications. Methods In the 24-center TRIUMPH registry, 4340 AMI patients underwent detailed interviews and 1631 reported smoking within 30 days of admission. Prescription of first-line smoking cessation medications at discharge was assessed by medical record review. All factors associated with smoking cessation treatment based on literature review were included in hierarchical modified log Poisson models. Results Only 14% (222/1631) of AMI patients who smoked were prescribed smoking cessation medication at discharge. Younger age, high school graduation, heavy cigarette usage, and treatment variables (instruction on smoking cessation, in-hospital revascularization) were independently associated with prescription of smoking cessation pharmacotherapy (see Figure). There was significant variation in prescribing across sites: range: 0-28%, median rate ratio 1.42 (1.24, 2.70) after multivariable adjustment. Conclusions Smokers surviving an AMI are rarely prescribed recommended smoking cessation treatment and there is considerable variation in prescription rates among hospitals. Although inpatient cessation counseling alone has not been shown to increase cessation rates in AMI patients, it increases the likelihood that patients receive effective pharmacotherapy. More aggressive use of pharmacotherapy is justified in older, less educated, and lighter smokers.

Practice-Level Variation in Warfarin Use Among Outpatients With Atrial Fibrillation (from the NCDR PINNACLE Program)

Warfarin is a complex but highly effective treatment for decreasing thromboembolic risk in atrial fibrillation (AF). We examined contemporary warfarin treatment rates in AF before the expected introduction of newer anticoagulants and extent of practice-level variation in warfarin use. Within the National Cardiovascular Data Registry Practice Innovation and Clinical Excellence program from July 2008 through December 2009, we identified 9,113 outpatients with AF from 20 sites who were at moderate to high risk for stroke (congestive heart failure, hypertension, age, diabetes, stroke score >1) and would be optimally treated with warfarin. Using hierarchical models, the extent of site-level variation was quantified with the median rate ratio, which can be interpreted as the likelihood that 2 random practices would differ in treating "identical" patients with warfarin. Overall rate of warfarin treatment was only 55.1% (5,018 of 9,913). Untreated patients and treated patients had mean congestive heart failure, hypertension, age, diabetes, stroke scores of 2.5 (p = 0.38) and similar rates of heart failure, hypertension, diabetes mellitus, and previous stroke, suggesting an almost "random" pattern of treatment. At the practice level, however, there was substantial variation in treatment ranging from 25% to 80% (interquartile range for practices 50 to 65), with a median rate ratio of 1.31 (1.22 to 1.55, p <0.001). In conclusion, within the Practice Innovation and Clinical Excellence registry, we found that warfarin treatment in AF was suboptimal, with large variations in treatment observed across practices. Our findings suggest important opportunities for practice-level improvement in stroke prevention for outpatients with AF and define a benchmark treatment rate before the introduction of newer anticoagulant agents.

Prevalence and correlates of smoking cessation pharmacotherapy in hospitalized smokers with acute myocardial infarction

Although current performance measures recommend smoking cessation counseling at the time of acute myocardial infarction (AMI), the American College of Cardiology/American Heart Association guidelines recommend pharmacotherapy as well. The aim of this study was to describe the prevalence and correlates of smoking cessation pharmacotherapy in hospitalized patients with AMI. In the 24-center TRIUMPH registry, 4,340 AMI patients underwent detailed interviews; and 1,631 reported smoking within 30 days of admission. Prescription of first-line smoking cessation medications at discharge was assessed by medical record review. All patient-related factors associated with smoking cessation treatment, based on literature review, were included in hierarchical modified log Poisson models. Only 14% (222/1,631) of AMI patients who smoked were prescribed smoking cessation medication at discharge. After multivariable adjustment for patient characteristics, there was significant variation across sites (range 0%-28%, median rate ratio 1.41, 95% CI 1.23-2.67). Independent factors associated with smoking cessation pharmacotherapy included older age (rate ratio 0.81 per 10-year increment, 95% CI 0.71-0.93), high school graduation (rate ratio 1.37, 95% CI 1.10-1.66), heavy cigarette usage (>20/d) (rate ratio 3.08, 95% CI 2.20-4.12), in-hospital revascularization (rate ratio 1.41, 95% CI 1.0-1.94), and instructions on smoking cessation (rate ratio 2.37, 95% CI 1.40-4.01). Smokers surviving an AMI are infrequently prescribed guideline-recommended smoking cessation treatments, and there is considerable variation across hospitals. Older, less educated, and lighter smokers are less likely to receive aggressive smoking cessation treatment. Novel strategies to augment current practice are needed.

Factors influencing patient willingness to participate in genetic research after a myocardial infarction

BackgroundAchieving 'personalized medicine' requires enrolling representative cohorts into genetic studies, but patient self-selection may introduce bias. We sought to identify characteristics associated with genetic consent in a myocardial infarction (MI) registry.MethodsWe assessed correlates of participation in the genetic sub-study of TRIUMPH, a prospective MI registry (n = 4,340) from 24 US hospitals between April 2005 and December 2008. Factors examined included extensive socio-demographics factors, clinical variables, and study site. Predictors of consent were identified using hierarchical modified Poisson regression, adjusting for study site. Variation in consent rates across hospitals were quantified by the median rate ratio (MRR).ResultsMost subjects consented to donation of their genetic material (n = 3,484; 80%). Participation rates varied greatly between sites, from 40% to 100%. After adjustment for confounding factors, the MRR for hospital was 1.22 (95% confidence interval (CI) 1.11 to 1.29). The only patient-level factors associated with consent were race (RR 0.93 for African Americans versus whites, 95% CI 0.88 to 0.99) and body mass index (RR 1.03 for BMI ≥ 25, 95% CI 1.01 to 1.06).ConclusionAmong patients with an MI there were notable differences in genetic consent by study site, but little association with patient-level factors. This suggests that variation in the way information is presented during recruitment, or other site factors, strongly influence patients' decision to participate in genetic studies.

Incidence, Correlates, and Outcomes of Acute, Hospital-Acquired Anemia in Patients With Acute Myocardial Infarction

Anemia is common among patients hospitalized with acute myocardial infarction and is associated with poor outcomes. Less is known about the incidence, correlates, and prognostic implications of acute, hospital-acquired anemia (HAA). We identified 2909 patients with acute myocardial infarction who had normal hemoglobin (Hgb) on admission in the multicenter TRIUMPH registry and defined HAA by criteria proposed by Beutler and Waalen. We used hierarchical Poisson regression to identify independent correlates of HAA and multivariable proportional hazards regression to identify the association of HAA with mortality and health status. At discharge, 1321 (45.4%) patients had HAA, of whom 348 (26.3%) developed moderate-severe HAA (Hgb <11 g/dL). The incidence of HAA varied significantly across hospitals (range, 33% to 69%; median rate ratio for HAA, 1.13; 95% confidence interval, 1.07 to 1.23, adjusting for patient characteristics). Although documented bleeding was more frequent with more severe HAA, fewer than half of the patients with moderate-severe HAA had any documented bleeding. Independent correlates of HAA included age, female sex, white race, chronic kidney disease, ST-segment elevation myocardial infarction, acute renal failure, use of glycoprotein IIb/IIIa inhibitors, in-hospital complications (cardiogenic shock, bleeding and bleeding severity), and length of stay. After adjustment for GRACE score and bleeding, patients with moderate-severe HAA had higher mortality rates (hazard ratio, 1.82; 95% confidence interval, 1.11 to 2.98 versus no HAA) and poorer health status at 1 year. HAA develops in nearly half of acute myocardial infarction hospitalizations among patients treated medically or with percutaneous coronary intervention, commonly in the absence of documented bleeding, and is associated with worse mortality and health status. Better understanding of how HAA can be prevented and whether its prevention can improve patient outcomes is needed.

Association of pulse wave velocity with vascular and valvular calcification in hemodialysis patients

The recent Kidney Disease: Improving Quality Outcomes (KDIGO) recommendations called for an investigation of the relationship between various radiological methods to assess cardiovascular calcification and measures of arterial stiffness. Accordingly, in 131 adult maintenance hemodialysis patients, we investigated the association of aortic pulse wave velocity (PWV) with calcification of cardiac valves on echocardiography, coronary artery, and thoracic aorta calcium on computed tomography and a calcification score of the abdominal aorta obtained on a plain abdominal X-ray. All tests were performed within a week. Mean PWV increased as the severity of coronary artery, thoracic, and abdominal aorta calcium scores increased (each P<0.05). No trend was present for number of valves with calcification. After multivariable adjustment, abdominal aorta X-ray calcium scores remained associated with PWV (P=0.004), whereas the association of PWV with thoracic aorta and coronary artery calcium scores became marginal (P=0.308 and P=0.083, respectively). No association was found between number of calcified valves and PWV. This study demonstrates a strong association between abdominal aorta calcification on plain X-ray and PWV and a borderline association with thoracic aorta and coronary artery calcification. Sudden death and congestive heart failure, two frequent causes of death in hemodialysis, are likely caused by increased arterial stiffness that can be closely predicted by the presence of aortic calcification on plain X-rays.

Implementation of Ergonomic Interventions in Healthcare: Results from 111 Facilities

Work-related injuries in healthcare facilities are highly prevalent. This study examines the effectiveness of ergonomic interventions in 111 healthcare facilities on reducing musculoskeletal disorders (MSDs). Data from OSHA logs were collected prior to (baseline) and after implementing (follow-up) engineering controls. The average length of the baseline period was 445 days, and the average length of follow-up was 629 days. Overall, 83 work units (64%) experienced a decrease in the MSD rates during the follow-up period (z = −5.193, p&lt; .001). The median rate ratio for all interventions was .63; that is, the intervention is associated with overall 37% decrease in MSD rates. The decrease in back injury rates observed in the study (44%) was much higher than the decline in the national data (17%). This study demonstrates that ergonomic interventions in healthcare facilities can result in decreased MSD rates.

Outcomes in older people undergoing operative intervention for colorectal cancer.

Outcomes in older people undergoing operative intervention for colorectal cancer.

P02 Adenosine-induced bronchoconstriction in the allergic rabbit: Evidence for A,-adenosine receptor

Among patients without cardiovascular disease (CVD) and low 10-year CVD risk, the risks of gastrointestinal bleeding and hemorrhagic strokes associated with aspirin use outweigh any potential atheroprotective benefit. According to the guidelines on primary prevention of CVD, aspirin use is considered appropriate only in patients with 10-year CVD risk ≥6% and inappropriate in patients with 10-year CVD risk <6%.The goal of this study was to examine the frequency and practice-level variation in inappropriate aspirin use for primary prevention in a large U.S. nationwide registry.Within the National Cardiovascular Disease Registry’s Practice Innovation and Clinical Excellence registry, we assessed 68,808 unique patients receiving aspirin for primary prevention from 119 U.S. practices. The frequency of inappropriate aspirin use was determined for primary prevention (aspirin use in those with 10-year CVD risk <6%). Using hierarchical regression models, the extent of practice-level variation using the median rate ratio (MRR) was assessed.Inappropriate aspirin use frequency was 11.6% (7,972 of 68,808) in the overall cohort. There was significant practice-level variation in inappropriate use (range 0% to 71.8%; median 10.1%; interquartile range 6.4%) for practices; adjusted MRR was 1.63 (95% confidence interval [CI]: 1.47 to 1.77). Results remained consistent after excluding 21,052 women age ≥65 years (inappropriate aspirin use 15.2%; median practice-level inappropriate aspirin use 13.8%; interquartile range 8.2%; adjusted MRR 1.61 [95% CI: 1.46 to 1.75]) and after excluding patients with diabetes (inappropriate aspirin use 13.9%; median practice-level inappropriate aspirin use 12.4%; interquartile range 7.6%; adjusted MRR 1.55 [95% CI: 1.41 to 1.67]).More than 1 in 10 patients in this national registry were receiving inappropriate aspirin therapy for primary prevention, with significant practice-level variations. Our findings suggest that there are important opportunities to improve evidence-based aspirin use for the primary prevention of CVD.